Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session A: Tuesday, August 12, 1:30 – 4:30 pm, de Brug & E‑Hall

Neural representations of implied motion in body-selective regions revealed by caption-based encoding models

Ryuto Yashiro¹, Masataka Sawayama², Ayumu Yamashita³, Kaoru Amano³; ¹Freie Universität Berlin, ²Hokkaido University, ³The University of Tokyo

Presenter: Ryuto Yashiro

Recent studies using encoding models that accurately predict brain responses to natural images from corresponding captions suggest that high-level visual regions encode the rich semantics of a scene, rather than merely responding to specific categories such as faces or bodies. Here we focused on the extrastriate body area (EBA) and investigated what semantic content is represented in this region by analyzing the relationship between the co-occurrence of multiple objects in large-scale natural scene captions and EBA responses predicted by caption-based encoding models. We found that responses in EBA as well as the fusiform body area (FBA) outside motion-processing regions exhibited high correlation with the perceived motion speed of a human body implied in an image, suggesting that these body-selective regions encode the implied motion of a human body inferred from the co-occurring objects in a scene.

Topic Area: Object Recognition & Visual Attention

Extended Abstract: Full Text PDF